But as private investors have been pushing ahead to develop the systems needed for these new robotic machines, one crucial innovator has been largely out of the loop: the United States military.

The armed forces have lagged on deploying their own versions of unmanned road vehicles, despite goals to create new machines that could be used in place of “boots on the ground” in conflicts. Restrictions on government spending and technological challenges have left the military with virtually no chance of meeting the goal set by Congress to have a third of the military’s combat fleet consist of unmanned vehicles by 2015, military experts said.

The military’s failure to lead the way in self-driving ground vehicles is ironic, given that today’s commercial advances have their roots in research originally sponsored by Darpa, the Defense Advanced Research Projects Agency, the Pentagon’s advanced technology organization. A decade ago, Darpa offered a series of “grand challenges” to private researchers, which helped push the technology forward.

Now General Motors and Nissan said last month they would offer self-driving cars to customers before the end of the decade. Early next year BMW and several other carmakers plan to offer more limited systems that will drive automatically in freeway traffic at low speeds. And Google already has a small fleet of vehicles with more than a half-million miles of automatic driving on California’s freeways.

“Now the automation of vehicles is taking off on the civilian side,” said Peter W. Singer, a Brookings Institution researcher and author of “Wired for War,” about the development of robot weapons. Mr. Singer predicts that civilian advances will ultimately trickle down to the military, a radical turnaround.

The military is not completely bereft of high-tech ground vehicles that can assist in warfare. The Legged Squad Support System, developed by Boston Dynamics, is a four-legged robot about the size of a cow. The system is intended to follow a soldier in the field, carrying up to 400 pounds of equipment.

However, the robot illustrates the technological challenges of making vehicles that not only have systems in place to complete tasks, but can operate and survive in unmapped, hostile environments.

“The hard problem in building autonomous ground vehicles is that the ground is hard,” said Gill Pratt, a program manager at Darpa.

Yet progress has been made. Some 1.5 million YouTube viewers marveled at the Legged Squad Support System’s ability to climb hills and walk over rocks. The designers’ real achievement, Dr. Pratt said, was in building a robot that wasn’t stumped by tall grass.

“One of the most important things you see in the videos is not the robot climbing up the hill,” Dr. Pratt said. “It was actually that it made the decision that it could go through the grass. I saw that and I said, ‘Wow!’ ”

Existing ground-based mobile robots are used largely for specialized military tasks. For instance, small tele-operated vehicles are used for clearing improvised explosive devices, known as I.E.D.’s, and for some surveillance tasks. In Iraq, about 8,000 tele-operated robots were used on about 125,000 missions.

In 2012, the first completely autonomous ground supply vehicle, built by Lockheed, was used to ferry more than 10,000 pounds of supplies to a combat outpost in Afghanistan that was more than a mile from a military base.

Next year, the military plans to have that vehicle, known as the Squad Mission Support System, delivered by a robotic helicopter to a location and controlled from several hundred miles away.

But such ground vehicles remain few and far between. In contrast, one-third of the military’s air fleet has been autonomous since 2012, meeting a goal set by Congress a dozen years ago.

At a recent military-oriented trade show at the Washington Convention Center, sleek Predator-style surveillance planes, robotic helicopters and hovering coffee table-size quad-copters could be spotted just about everywhere. But only a handful of unmanned ground systems were shown, and they were based on technology half a decade old.

The imbalance between air and land systems can be seen in Pentagon spending. The budget ending Sept. 30 allocated $6.04 billion for autonomous aircraft and just $261 million for unmanned ground vehicles.

The gap, said John Arquilla, a military strategist at the Naval Postgraduate School in Monterey, Calif., a research university operated by the Navy, is “particularly troubling because a large percentage of our casualties were people driving vehicles blown up by I.E.D.’s.” If trucks in Afghanistan and Iraq had been robotic, he said, “casualties would have been cut by two-thirds over the last decade.”

Maureen Schumann, a spokeswoman for the Office of the Secretary of Defense, said the goal of deploying autonomous ground systems was taken off the table when Congress killed the financing for the Future Combat System in 2009. The system, a $340 billion project, was an ambitious effort to modernize the Army with both manned and unmanned vehicles. Critics said spending for it had run amok, and the goals did not necessarily address the challenges posed by today’s terrorists and insurgent forces.

“The collapse of F.C.S. has knocked the Army out of the technology business,” said James Lewis, director of the technology and public policy program at the Center for Strategic and International Studies, a bipartisan research and analysis group in Washington. “They are now really focused on helicopters and less on the kind of vehicles used for humping stuff around the ground.”

In a 2012 report, the Pentagon’s Defense Science Board, an advisory group, noted public fears that the move toward robot warriors would rapidly displace human judgment in warfare. Questions have also been raised about the morality — and strategic effectiveness — of their use.

The military, too, has its reservations about robots usurping the traditional soldier’s role on the ground.

“Over time we are slowly knocking down this wall, but there is a resistance to new technologies being introduced in and around soldiers,” said Don Nimblett, a senior manager for Lockheed Martin’s combat maneuver systems, a contractor that could benefit from Pentagon spending on such systems. “We’ve gone about as far as we can. At some point the government has to make it into a program and fund it.” 

Source:The New York Times